My goal is to argue that the absence of a methodology of grammaticality judgements in linguistics constitutes a serious obstacle to meaningful research, and to begin to propose a suitable remedy. Since at least the beginning of the generative paradigm in linguistics, judgements of the grammaticality / acceptability of sentences have been the major source of evidence in constructing grammars, leading some to suggest that theoretical linguists are in fact constructing grammars of linguistic intuitions, which need not be identical with the competence underlying production or comprehension. Also, in this pseudo-experimental procedure of judgement elicitation, there is typically no attempt to impose any of the standard experimental control techniques, and often the only subject is the theorist himself or herself. We provide a survey of how grammaticality judgements are currently used in theoretical syntax, and argue that such uses, in combination with the problems of intuition and experimental design, demand a careful examination of judgements, not as pure sources of data, but as instances of metalinguistic.
Several important issues arise when this view of grammaticality judgements is taken, including what tasks one can use to elicit them, how people might go about giving them, and what they might tell us about linguistic competence. Our central hypothesis is that grammaticality judgements result from interactions between primary language faculties of the mind and general cognitive properties, and crucially do not involve special components dedicated to linguistic intuition. We review the psycholinguistic research that has examined ways in which the judgement process can vary with differences between subjects and with experimental manipulations. Parallels with other cognitive behaviours that our hypothesis predicts are pointed out. We then integrate the substantive and methodological findings in the form of a model of linguistic knowledge that reflects what is known about linguistic intuitions, and a proposed methodology for collecting grammaticality judgements while avoiding the pitfalls of previous work and taking account of the conditions that have been shown to influence them. Finally, we discuss how mainstream linguistic theory might be affected by the growing body of research in this area.
" }